<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"><!-- saved from url=(0035)file://C:\WINDOWS\TEMP\UNILEO4.html --><!-- saved from url=(0022)http://internet.e-mail -->

<small>DRAFT # 5: UPDATED 01.08.2000</small>

Simultaneous Measurements of Temperature
in the Upper Mesosphere with an Ebert-Fastie
Spectrometer and a VHF Meteor Radar on
Svalbard (78°N, 16°E)

K. P. Nielsen1, J. Röttger2 and F. Sigernes3

1 The University Courses on Svalbard and The University of Bergen, Norway.

2 The Max-Planck-Institute of Aeronomy, Germany.

3 The University Courses on Svalbard.


From the 14th of November to the 21st of November 1999 measurements of the upper mesospheric temperatures were performed. Both the SOUSY Svalbard Radar and a one meter focal length Ebert-Fastie spectrometer were used, the first measuring the decay of meteor echoes and the latter measuring the OH(6-2) rotational spectrum. Due to the differences in the height of the measurements the temperatures were compared on the background of the temperature profile according to the COSPAR International Reference Atmosphere and reasonable agreement was found. During good conditions for the optical measurements the rotational temperatures had a relative uncertainty of 2 Kelvin when one-hour averages were taken. A frequency analysis of these temperatures was performed, showing no signs of neither diurnal nor semi-diurnal variations. Suggesting that these are not present or weaker than 2 Kelvin in the OH-airglow layer over Svalbard in November.


<big>1. Introduction</big>

Temperature profiles in the mesopause region at high polar latitudes of 80 degrees are fairly rare (Connor, 1993). The highest latitudes, where temperature profiles were measured in the past, are about 70 degrees (Lübken, 1999). We do not know of any direct comparisons of temperature measurements at such latitudes. The unique location of Svalbard at 80 degrees N, where many scientific observations of the middle and upper atmosphere are done (Bjørdal, 1998), would allow such observations and comparisons. For this purpose we have performed measurements of the upper mesospheric temperatures in the period of from the 14th to the 21st of November 1999. Both the SOUSY-Svalbard-Radar (Röttger, 2000) and an Ebert-Fastie spectrometer (Dick et al., 1970) were used.

Observations of tidal variations in the Mesosphere-Lower Thermosphere region over Svalbard have been done with the EISCAT Svalbard Radar (Aso et al., 1999 and Van Eyken et al. 1999). These were only of neutral winds not of temperatures, and only during summer. Winter observations of tides have been performed earlier with the Ebert-Fastie spectrometer (e.g. Myrabo and Deehr, 1984), but not in November. Myrabo and Deehr found a semi-diurnal tide with an amplitude of 3 Kelvin in December 1982.

The spectrometer is placed at the Auroral Station that is situated in Adventdalen, close to Longyearbyen on Svalbard. The radar is practically situated right next to the Auroral Station, as they are only a few kilometers apart. The radar can obtain temperatures through measurements of the echoes from trails of ionization resulting from ablating meteorites. These echoes decay by diffusion; as the diffusion co-efficient is a function of temperature and pressure, the temperature can be estimated. However, uncertainty remains, since these estimates require information of the pressure, which needs to be taken from a model (Hocking et al., 1999). The spectrometer measures the OH(6-2) rotational-vibrational spectrum; from its emissions the temperature was calculated by the aid of synthetic spectra. The emissions originate from the OH-airglow-layer that is situated at an average height of 87 km, while the meteor echoes occur at a variety of heights in the upper mesosphere and lower thermosphere. We present here initial comparisons. Further detailed analysis should follow as well as more joint observations.


<big>2. Instrumentation</big>

The primary aim of our joint effort was to compare two methods of measuring the temperatures in the upper mesosphere. The temperatures were obtained using the SOUSY-Svalbard Radar (SSR) running in the meteor-radar-mode and the one-meter focal length Ebert-Fastie spectrometer measuring the airglow hydroxyl rotational temperature. As the SSR needs a significant number of meteor echoes in order to estimate the temperature, the campaign was carried out at the time of the Leonid meteor shower. Being a project of UNIS (the University Courses on Svalbard), the campaign was called UNILEO (UNIS and SOUSY observations around the Leonid shower).

Both the SSR and the spectrometer were set up to measure around zenith with fields of view of 4.5 and 5.0 degrees, respectively. The radar has the possibility to point its beam to the NE, SE, SW and NW with 5 degrees zenith angles and the spectrometer measures around the zenith direction. Where the spectrometer only measures the temperature as an average through the OH(6,2)-airglow layer (9km x 12km at 87 km), the SSR measures temperatures in whichever height incoming meteors ablate. The OH-airglow layer is, according to in-situ measurements made with rockets, situated around 87km with a thickness of 8km regardless of season or latitude (Baker and Stair 1988).

2.1 The Ebert-Fastie Spectrometer

The Ebert-Fastie spectrometer is one out of three spectrometers located at the Auroral Station in Adventdalen. The instrument has a scanning grating that scans through the first order wavelength range 8312 - 8745Å. The bandpass is 4.8Å and the field of view is approximately 5°. For a more detailed description of the spectrometer see Fastie (1952), Dick et al. (1970) and Sivjee et al. (1972). In order to derive the rotational temperatures of the measured OH(6,2)-band, it is necessary to produce the corresponding synthetic spectra. The basis for calculating these spectra is given by Herzberg (1950) with term values from Krassovsky (1962) and Einstein coefficients from Turnbull and Lowe (1989). The temperatures can then be derived since the upper populated energy states of the OH(6,2) band are distributed according to the Stefan-Boltzmann distribution. Recent work by Conner (1993) explains in detail how the parameters necessary for generating the synthetic spectra as a function of the rotational temperature can be calculated. The temperature is found by choosing the optimal fit between the measured and the synthetic spectrum through iteration, to minimize the least square error.

Fig. 1 shows a typical result of fitting a synthetic spectrum to the data obtained during the UNILEO campaign. In this case, a temperature of 204°K with a fit value of 0.95 was found. The fit value is defined as

,                                           (1)

where l is the least mean square difference between the measured and the synthetic spectrum. This is the same fit function that was used by Viereck (1988). A Fit=0.8 represents a relative uncertainty of approximately ± 2°K.

2.2 The SOUSY Svalbard radar

The SOUSY Svalbard Radar (SSR) of the Max-Planck-Institut für Aeronomie (MPAe) is an MST radar (Röttger, 2000). It uses the main basic components of the mobile SOUSY radar, which was constructed near Longyearbyen in summer 1998. It consists of a transmitter operating on 53.5 MHz at peak power of 150 kW (60 kW with 2.5 kW average power during this campaign), and a phased-array antenna system of 356 Yagi antennas (total 33 dBi gain and 4 degrees half-power beamwidth). Phase steering allows five beam pointing directions at zenith and 5 degrees zenith angle to NE, SE, SW and NW. Transmitter, radar control, receiver and digital signal processing units and auxiliary equipment is housed in three containers. For the meteor observations a 20-bit complementary code with 2 us baud length is used, corresponding to 300m range resolution.

 The SSR was not originally conceived or optimized as meteor radar, but it detects meteors as any other MST radar. This can be used to get wind and temperature estimates, and some information about the entrance and ablation of meteorites in the lower thermosphere and upper mesosphere. For comparison with the OH spectrometer observations, around the Leonid meteor shower in November 1999, the SSR was operated in a special meteor radar mode with a short coherent integration sampling time of 28 ms and antenna beam directions fixed for longer periods. The system did not allow a faster data rate, which would have been a more suitable choice. The antenna beam direction was kept constant for about 1hour and then switched to the orthogonal direction for another hour. More than 4 days of meteor radar data were collected. We concentrate here on underdense echoes, where the echo amplitude decay can be used to deduce the diffusion coefficient and hence the temperature (e.g., Hocking, 1999). Using equation (4) of Hocking et al. (1997) we obtain the temperatures 

,                                       (K)     (2)

where p is the pressure in Pascal and t1/2 the time at which the amplitude of the meteor echo has fallen to half its peak value. The pressure was taken from the CIRA 1986 model. For a geometric height of 84.8 km this is 48.5 Pa.

Fig. 2 shows an example of underdense meteor echoes observed in the evening before the Leonid shower maximum. These prove the exponential decay of the echoes. We use these kinds of echoes to determine decay time and diffusion coefficient and hence the temperature can be estimated using equation (2).


<big>3. Results</big>

Fig. 3 shows the temperature estimates obtained from the spectrometer as a function of time throughout the campaign The time axis is day number in 1999. Each cross on the figure represents the mean temperature of one hour of measurements. As it can be seen the data is spread over a rather large temperature interval ranging over more than 50 K around an average value of 200-210°K, illustrating the large temperature variations of the upper mesosphere. The time of the Leonid meteor shower on the morning of the 18th of November is marked on the figure. The crosses with boxes represent data that have a fit function (equation (1)), which is better than 80%. These temperature estimates are compared with simultaneously deduced temperature estimates obtained from the radar.

As an initial attempt we have carefully screened the SSR observations on 18 November 1999, 0000 UT +-3 hours. The latter part of this period covers the Leonid shower when about three times as many meteor echoes were observed than usual. In total we found some 35 meteor echo events in the altitude range 82-90 km which fulfilled our criteria of sufficient signal-to-noise ratio and proper exponential decay. This relatively low rate of meteor events results from the fact that the antenna beam is just 4 degrees wide. However, it was possible to obtain estimates of the diffusion coefficient and the temperature, which we classified to cover height ranges 82-86 km and 86-90 km.

Fig. 4 shows the CIRA temperature profile and superimposed the temperature estimates for the given time period. The variance of measured temperatures is fairly broad for both instruments, due to measurement uncertainties and geophysical noise (probably due to shorter period gravity waves). The mean value, however, matches quite well with the model temperature profile.


<big>4. Discussion</big>

There are some matters of uncertainty to the accuracy of the OH-rotational temperatures. Pendleton et al. 1989 questioned whether excited OH is in thermal equilibrium. They made spectral measurements of the (7,4)-band and came to the conclusion that for rotational levels above J´ = 5 this is not the case. In general whether excited OH is in thermal equilibrium or not is definitely a question of vibrational level, as the lifetime of the molecule decreases when the vibrational level increases. The collision frequency in this altitude of the atmosphere is of the order 104s-1, while the average lifetime of OH in vibrational state 6 is of the order 10-2-10-3s-1 (Turnbull and Lowe, 1989). Thus it is reliable that OH(v=6) is in thermal equilibrium for the lower rotational states. The temperature derived from the OH-airglow therefore is equal to the average temperature of the surrounding atmosphere with a standard error of below 10K. Another matter of uncertainty is the height and thickness of the airglow layer. Baker and Stair (1988) compared previous measurements of the layer made from rockets and found that the height is nearly constant regardless of latitude or season. Still this is the biggest problem with the OH-temperatures. In particular as only few measurements of the OH-height has been performed at high latitudes. As there are strong temperature gradients in this part of the mesopause region, and the airglow temperature results from the OH-layer 3 kilometers higher up, the observed difference between the temperatures could be natural. However, comparison of more measurements is needed to investigate this.

The temperature estimates obtained from the radar are quite sensitive to the time t1/2 and the pressure. Whereas p/t1/2 can be obtained with acceptable accuracy by choosing a significant amount of proper underdense meteor echoes, the pressure cannot directly be measured but has to be taken from a model. Hocking et al. (1999) have tried to untangle this problem of pressure dependence. The large variance of the data points, as we observe with the radar, is similar to observations at other locations (Chilson et al., 1996; Hocking et al., 1999). It partially results from influence of noise on the signal, inhomogeneity of the meteor trail during diffusion, or on the ionic composition of the ablating meteorite.

A comparison of the measured temperatures and the CIRA 1986 average temperatures for November was made (figure 4). As it can be seen the temperatures are made in different heights and the difference between them is corresponding fairly well with the gradient in the temperatures from the CIRA model. It is also seen that the mean OH rotational temperatures are more than 10K below the model temperatures.

A direct measure of the pressure can be carried out by using the temperatures from the spectrometer combined with the radar measurements of t1/2. The method may be a good candidate for improving model calculations. This should be performed later. These preliminary observations prove that useful information can be extracted from the meteor observations of the SSR. We think, however, that a special purpose meteor radar with wider beam and receiving interferometry would be an advisable improvement for such radar observations (Röttger, 2000). The combination with the OH spectrometer observations of temperature is a most useful improvement.

In figure 5 an example of a Lomb normalized periodogram of the OH rotational temperatures is shown. The spectrum has no peaks stronger than 5 Kelvin, which is only slightly more than the uncertainty of the temperatures (2 Kelvin). This indicates that no tides were present over Svalbard in November 1999, which had stronger amplitude than 2 Kelvin. This differs from the neutral wind measurements of Aso et al. (1999). Their measurements were done with the EISCAT Svalbard radar also located in Adventdalen. They found that the diurnal and semidiurnal tides were the dominating modes in the region between 90 km and 120 km height and of considerable strength. However their measurements were done in August, not November. Myrabo and Deehr (1984) found semi-diurnal variations of the temperatures in December with an amplitude of 3 Kelvin, which if they where present, just barely would have been detected from the OH rotational temperatures. Meaning that this does not contradict the present measurements. Also the results of the GSWM-98 (Hagan et al., 1999) are in agreement with the present measurements, as the amplitudes of both the diurnal and semi-diurnal tides in this are 1 Kelvin or weaker at OH-airglow heights over Svalbard.


<big>5. Concluding remarks</big>

The temperatures measured with the Ebert Fastie spectrometer and the SSR were in reasonably good agreement. The simultaneous measurements may give rise to later pressure measurements in the height of the airglow layer. We will look into this in the future. However, we have to be aware that such observations can yield reliable estimates only if long averages are possible. Another result of this campaign was that neither diurnal nor semi-diurnal variations of temperature were detected.


<big>Figure Captions</big>

Fig. 1. The measured and synthetic spectra of the OH(6,2) band. Each line is marked and identified according to quantum state. The curve plotted with line color red is the synthetic, green the measured and yellow is the detected background spectra, respectively. The spectral lines plotted with blue color represents emissions caused by aurora.

Fig. 2. Underdense meteor echoes observed with the SOUSY-VHF-Radar on 17.11.1999 at altitudes 82.6 - 85.3 km. The antenna was pointing to NW at elevation 85°. The record covers 1.8 seconds. A single sample was obtained every 28 milliseconds. The pressure at 85 km height is according to the CIRA86 model 4.9 Pa. Thus the echo decay rate corresponds to a temperature of about 220 Kelvin (from equation (2)).

Fig. 3. Time series of airglow hydroxyl temperatures obtained by the Silver Bullet spectrometer at the Auroral Station in Adventdalen during the UNILEO campaign, 14-21 November 1999. Each point represents an hourly averaged temperature. Square and crossed symbols represent fit values above or below 80%, respectively.

Fig. 4. The zonal mean upper mesospheric temperatures in November at 80°N as a function of height according the COSPAR International Reference Atmosphere model from 1986 (CIRA 1986). On the figure the single meteor-radar temperature and the OH-rotational temperatures are marked. The error bar of the meteor temperature is 10° K. The temperatures from the spectrometer correspond to the spread of the temperatures throughout the campaign.

Fig. 5. A Lomb-Scargle analysis of the OH-rotational temperatures. The analysis is made of 1-hour average temperatures from the period between the 13th of November and the 30th of November 1999. 



We appreciate the support of this campaign by the University Courses on Svalbard (UNIS).



Andrews, D. G., J. R. Holton and C. B. Leovy, Middle Atmosphere Dynamics, Academic Press, Orlando, Florida, 1987.

Aso, T., S. Ito and S. Kato, Background Wind Effect on the Diurnal Tide in the Middle Atmosphere, Jour. Geophys. Res., vol. 86, 11.388-11.400, 1987.

Aso, T., A. van Eyken, and P. J. S. Williams, EISCAT Svalbard Radar derived atmospheric tidal features in the lower thermosphere as compared with the numerical model ATM2, Adv. Upper Atmos. Res., 13, 48-56, 1999.

Baker, D. J and A. T. Stair, Jr., Rocket Measurements of the Altitude Distributions of the Hydroxyl Airglow, Physica Scripta. Vol. 37, 611-622, 1988.

Bjørdal, A., Research in Svalbard 1998, Norsk Polarinstitutt, Tromsø, 1998.

Chilson, P. B., P. Czechowsky, and G. Schmidt, A comparison of ambiploar diffusion coefficient in meteor trains using VHF radar and UV lidar, Geophys. Res. Lett., 23, 2745-2748, 1996.

Conner, J., An Investigation of the Dynamics of the Mesopause. Fabry-Perot Observations of Winds and Temperatures from Nightglow Emissions. Ph. d. thesis., Fairbanks, 1993.

Dick, K. A., G. G. Sivjee and H. M. Crosswhite, Aircraft Airglow Intensity Measurements: Variations in OH and OI (5577), Planet. Space Sci., Vol. 18, 887, 1970.

Fastie, W. G., A small plane grating monochromator, J. Opt. Soc. Am., Vol. 42, 641, 1952.

Fastie, W. G., Image forming properties of the Ebert monochromator, J. Opt. Soc. Am., Vol. 42, 647, 1952.

Hagan, M. E., M. D. Burrage, J. M. Forbes, J. Hackney, W. J. Randel and X. Zhang, GSWM-98: Results for Migrating Solar Tides, J. Geophys. Res., Vol. 104, No. A4, 6813-6827, 1999.

Herzberg, G., Spectra of Diatomic Molecules, D. Van Nostrad co., inc. New Y.,1950.

Hocking, W. K., T, Thayaparan, and J. Jones, Meteor decay times and their use in determining a diagnostic mesospheric temperature-pressure parameter: methodology and one year of data, Geophys. Res. Lett., Vol. 23, 2977-2980, 1997.

Krassovsky, V. I., N. N. Shefov and V. I. Yarin, Atlas of the Airglow Spectrum 3000-12,400 Å, Planet. Space Sci., vol., 1962.

Pendelton, W, P. Espy, D. Baker, A. Steed, M. Fetrow and K. Henriksen, Observation of OH Meinel (7,4) P(N´´ = 13) Transitions in the Night Airglow, Jour. Geophys. Res. Vol. 94, No. A1, 505-510, 1989.

Lübken, F.-J., Thermal structure of the Arctic summer mesosphere, J. Geophys. Res., 104, 9135-9149, 1999.

Myrabo and Deehr, Tides in OH airglow over Svalbard, Planet Space Sci., 32, 263-271, 1984.

Röttger, J., Radar investigations of the mesosphere, stratosphere and the troposphere in Svalbard, Adv. Polar Upper Atmosph. Res., in print, 2000.

Sivjee, G. G., K. A. Dick, and P. D. Feldman, Temporal Variations in the Night-Time Hydroxyl Rotational Temperature, Planet. Space Sci., Vol. 20, 261, 1972.

Turnbull, D. N. and R. P. Lowe, New Hydroxyl Transition Probabilities and their Importance in Airglow Studies, Planet. Space Sci., Vol. 37, No. 6, 723-738,1989.

Van Eyken, A. P., P. J. S. Williams, M. Kunitake and S. Buchert, First E-region observations of tides in the lower thermosphere using the EISCAT Svalbard Radar. Submitted to Geophys. Res. Lett., 1999.

Viereck, R., Effects of Gravity Waves on the Polar O2 & OH Airglow. Thesis work, Fairbanks 1988.